jojo wrote:
> GL_LUMINANCE is for gray scale images. For example if you want to save the
> depth buffer.
Ok, thus, I should use GL_RGBA.
jojo wrote:
>
> Besides you should post a little bit more code. I don't know how much
> experience your have with open scene graph. Maybe you didn't call the frame()
> method of your viewer, the camera is not attached to a node or maybe there is
> just nothing to see in the viewing frustum?
>
My camera is attached to the root of the scene graph.
I retrieve the image during the rendering loop (it's true I didn't do it at
first...), thus after 5 viewer.frame ('cause I setted it at 5).
There is something to see from my camera.
jojo wrote:
>
> Did you display the camera normaly which means with turned off offscreen
> rendering?
>
However, I don't know what you mean here, it seems to be a key point...
I've read things about pre-render, but you don't do any, right?
Here's a bit of my code:
I create the camera:
Code:
ref_ptr<Camera> MyPTMutils::createCamera(Matrix& cameraIntrinsic, Matrix&
cameraPosition)
{
// parameters for MyPTM
double width = cameraIntrinsic(0, 2)*2;
double height = cameraIntrinsic(1, 2)*2;
double fy = cameraIntrinsic(1, 1);
// create camera for MyPTM
ref_ptr<Camera> new_camera = new Camera();
new_camera->setComputeNearFarMode(osg::CullSettings::DO_NOT_COMPUTE_NEAR_FAR);
new_camera->setViewport(0, 0, width, height);
new_camera->setProjectionMatrixAsPerspective((360./PI)*atan(height/(2*fy)) ,
(double)width/height, 1.0, 1000.0);
new_camera->setViewMatrix(Matrix::inverse(cameraPosition));
return (new_camera);
}
I modify the camera (I know this can seem strange out of the context):
Code:
MyPTM::MyPTM(Geode* _arg_geode, std::string imgpath , int _arg_textureUnit,
float focal, Quat &rotation, Vec3d position)
{
ref_ptr<osg::Image> img = osgDB::readImageFile(imgpath);
cu = img->s();
cv = img->t();
Vec2f *vec = new Vec2f(cu,cv);
MyPTMutils::setClosestPowerOfTwo(vec);
...
_camera = MyPTMutils::createCamera(cameraIntrinsic, transform);
_texture = new Texture2D(osgDB::readImageFile(imgpath));
_camera->setReferenceFrame(Transform::ABSOLUTE_RF);
_camera->setDataVariance(Object::DYNAMIC);
_depth = new Image();
_depth->allocateImage(vec->x(), vec->y(), 1, GL_RGBA, GL_UNSIGNED_BYTE);
_camera->attach(osg::Camera::COLOR_BUFFER, _depth.get(),0,0);
And the print:
Code:
float* bl = (float*)(_depth->data());
osgDB::writeImageFile(*_depth.get(),"depth.png");
This makes me think: why a float* ?
And the main
Code:
MyPTM* myptm2 = new MyPTM((Geode*) (root->getChild(2)),
"../Media/Data_Calibrage/Image/IMG_5316_small.jpg", 0, 720.0, r, camposition);
myptm2->_camera->setView(&viewer);
viewer.addSlave(myptm2->_camera.get());
viewer.setSceneData( root.get());
viewer.realize();
int once = 0;
while( !viewer.done() )
{
viewer.frame();
if (once == 50)
myptm2->printDepthImage();
once++;
}
------------------
Read this topic online here:
http://forum.openscenegraph.org/viewtopic.php?p=12271#12271
_______________________________________________
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
